VOLUME 119, NUMBER 23 JUNE 11, 1997 © Copyright 1997 by the American Chemical Society
Electric Field Effects on Electron Transfer Rates in Dichromophoric Peptides: The Effect of Helix Unfolding Marye Anne Fox* and Elena Galoppini Contribution from the Department of Chemistry and Biochemistry, UniVersity of Texas at Austin, Austin, Texas 78712 ReceiVed September 17, 1996X
Abstract: The effect of helix unfolding on the rates of photoinduced electron transfer in model dichromophoric peptides was investigated. Two R-helical peptides, 1 and 2, having an alternating Ala-Aib backbone and differing only in the position of an appended electron donor (N,N-dimethylaniline) and an appended photoexcited electron acceptor (pyrene) relative to the electric field generated by the helix, had shown a difference in photoinduced electron transfer rates which had been ascribed to a helix dipole effect. Upon denaturation by protic solvents (EtOH, MeOH, H2O, CF3CH2OH) or guanidinium, the observed electron transfer rates in 1 and 2 became identical. The helix unfolding was studied by circular dichroism (CD) analysis. A second pair of model oligopeptides, 3 and 4, analogous to 1 and 2 but having L-proline (Pro) instead of R-methylalanine (Aib) incorporated into the backbone, were prepared in order to study unfolded peptides in low dielectric constant solvents. The CD, NMR, and steady-state fluorescence spectra in a variety of solvents establish that one of the chromophores experiences a different local environment in 3 than in 4 and that the two peptides have different average conformations.
Introduction In an R-helical oligopeptide, the dipole moments of the amino acid residues align in the same direction, nearly parallel to the helix axis. The resulting macroscopic dipole generates an electrostatic potential, directed from the N-terminus to the carboxy-terminus, Figure 1.1 This electrostatic field, estimated to be up to 109 V/m in an R-helix,1b,2 plays an important role in the structure and functions of proteins, and probably also influences the primary electron transfer event in photosynthesis. It has been proposed, for example, that the permanent field of the first layer of proteins that surrounds the photosynthetic reaction center promotes a rapid electron transfer over a X Abstract published in AdVance ACS Abstracts, May 15, 1997. (1) For a review on the role of an R-helix dipole on protein function and structure see: (a) Hol, W. G. J. Prog. Biophys. Mol. Biol. 1985, 45, 149. (b) Hol, W. G. J.; van Duijnen, P. T.; Berendsen, H. J. C. Nature 1978, 273, 443. (c) Wada, A. AdV. Biophys. 1976, 9, 1. (d) Abraham, R. J.; Hudson, B. D.; Thomas, W. A.; Krohn, A. J. Mol. Graph. 1986, 4, 28. (2) The reported value of 109 V/m is calculated from vacuum electrostatics.
S0002-7863(96)03269-6 CCC: $14.00
Figure 1. Schematic drawing of the postulate orientation of the chargeseparated pair in a photosynthetic center embedded in parallel helices.
relatively long distance, producing a charge-separated pair oriented antiparallel to the helices’ electric field.1a This effect has been the object of both theoretical and experimental investigations.1a,3 (3) (a) Gosztola, D.; Yamada, H.; Wasielewski, M. R. J. Am. Chem. Soc. 1995, 117, 2041. (b) Brothers, H. M., III; Zhou, J. S.; Kostic, N. M. J. Inorg. Organomet. Polym. 1993, 3, 59. (c) Lockhart, D. J.; Kim, P. S. Science 1992, 257, 947. (d) Franzen, S.; Lao, K.; Boxer, S. G.; Chem. Phys.
© 1997 American Chemical Society
5278 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Fox and Galoppini
In previous work, we reported the photophysics of R-helical model peptides 1 and 2, showing that the rates of electron transfer between a donor-acceptor pair are indeed significantly influenced by the orientation of the charge-separated ion pair with or against the helix dipole.4,5 These two peptides differ
only in the positions of the N,N-dimethyl-p-anilino (the electron donor) and 2-pyrenyl (the electron acceptor) groups with respect to the helix ends. The calculated dipole moment for each peptide is approximately 40 D.4b From analysis of measurements of time-resolved pyrene fluorescence, the rates of electron transfer were shown to be 5-27 times faster in 1 (k1) than in 2 (k2), with smaller ratios being observed in solvents of higher dielectric constant.4a This result is in agreement with the postulated effect of the helix dipole, because faster rates were observed in 1, where the radical ion pair is antiparallel to the helix field. However, we consistently observed biexponential decays in 1 and 2, a clear indication of multiple conformations of the oligopeptides in the ground state, probably caused by changes in the torsional angle of the methylenic unit that links each chromophore to the backbone.6 Extended MM2-force field model calculations and semiempirical single-point calculations suggested that, although the peptide backbone is rigid, the side chains do experience sufficient conformational mobility4b to restrict the precision of the interpretation of the observed difference in electron transfer rates. To understand the relative importance of different conformations and the helix electric field, we have studied the effect of denaturation of 1 and 2 on the observed electron transfer rates. If the observed difference in rates is caused mainly by the helix dipole, the difference in electron transfer rates in denatured 1 and 2 should disappear, as the net dipole of a random coil peptide is zero. Lett. 1992, 197, 380. (e) Alegria, G.; Dutton, P. L. Biochim. Biophys. Acta 1991, 234, 1057 and 258. (f) Warshel, A.; A° qvist, J. Annu. ReV. Biophys. Chem. 1991, 20, 267. (g) Boxer, S. G.; Lockhart, D. J.; Franzen, S. In Photochemical Energy ConVersion; Norris, J. R., Meisel, D., Eds.; Elsevier: Amsterdam, 1989; p 196. (h) Boxer, S. G.; Goldstein, R. A.; Franzen, S. In Photoinduced Electron Transfer Part B. Experimental Techniques and Medium Effect; Fox, M. A., Chanon, M., Eds.; Elsevier: Amsterdam, 1988; p 163. (i) Abraham, R. J.; Hudson, B. D.; Thomas, W. A.; Krohn, A. J. Mol. Graph. 1986, 4, 28. (4) (a) Galoppini, E.; Fox, M. A. J. Am. Chem. Soc. 1996, 118, 2299. This reference describes the synthesis of 1 and 2 in Supporting Information. (b) Knorr, A.; Galoppini, E.; Fox, M. A. J. Phys. Org. Chem., in press. The calculated value of 40 D is probably overstimated, since the calculation does not take into account a saturation effect that occurs with increasing number of residues.1 (5) Evidence of the formation of a radical ion pair was provided by observation of the pyrene radical anion in the transient absorption spectra of 1 and 2.4b (6) Short (5-15 amino acids) peptides can adopt rapidly interconverting conformations in solution, and biexponential decays were often observed in studies of electron transfer in peptides: (a) Pispisa, B.; Venanzi, M.; Palleschi, A.; Zanotti, G. J. Phys. Chem. 1996, 100, 6835. (b) Pispisa, B.; Venanzi, M.; Palleschi, A.; Zanotti, G. Macromolecules 1994, 27, 7800. (c) Inai, Y.; Sisido, M.; Imanishi, Y. J. Phys. Chem. 1991, 95, 3847. (d) Inai, Y.; Sisido, M.; Imanishi, Y. J. Phys. Chem. 1990, 94, 6237. (e) Basu, G.; Kubasik, M.; Anglos, D.; Secor, B.; Kuki, A. J. Am. Chem. Soc. 1990, 112, 9410.
Figure 2. Circular dichroism spectra of 2 (1 × 10-4 M solutions in CH3CN (a); MeOH (b); CH3CN/H2O (c); CF3CH2OH (d)).
We report here the effect of denaturation on photoinduced electron transfer between appended donor- and acceptorsubstituted peptides, achieved by changing the external environment, such as solvents or salts, and by incorporation of proline into the oligopeptide sequence.7 Thermal denaturation was not employed in order to avoid ambiguities caused by partial chemical decomposition.8 Results and Discussion Denaturation of Peptides 1 and 2. Protic solvents and guanidinium salts often unfold peptides that exist as helices in organic solvents by promoting the formation of intermolecular hydrogen bonding, thus disrupting the intramolecular hydrogen bonding which leads to the helix. This effect is shown by the circular dichroism (CD) spectra of 2 in protic solvents, Figure 2. Whereas the CD spectrum of 2 in acetonitrile shows a strong positive band at 190 nm and two negative bands near 210 and 220 nm, which are characteristic of a right-handed helical conformation,7,9 the spectra in MeOH, CH3CN/H2O (70/30 (v/ v)), and CF3CH2OH provide evidence that the helical structure is partially retained in methanol and that water and 2,2,2trifluoroethanol disrupt the helix. An analogous observation was made for peptide 1. The CD spectrum of 1 in the presence of aqueous guanidinium carbonate, Figure 3, indicates a strong perturbation of the continuity of the helical structure, with possibly local helical subunits but not overall helicity (cf. Figure 6). However, it is unclear whether the major effect is exerted by guanidinium or by water. Time-Resolved Fluorescence Studies of 1 and 2 in Protic Solvents. The electron transfer rates of unfolded 1 and 2 were studied by single photon counting. Pyrene fluorescence lifetimes in the denatured peptides, Table 1, were compared with the values obtained in solvents in which 1 and 2 are helical. When guanidinum carbonate was added to THF solutions of 1 and 2, the ratio between the electron transfer rates of 1 and 2, (7) For a review about helix conformations, see: Goodman, M.; Verdini, A.; Choi, N. S; Masuda, Y. Top. Stereochem. 1970, 5, 69. Methods for denaturation of helical peptides are used or mentioned by Lockhart and Kim.3c (8) We observed that 1 and 2 rather easily undergo chemical decomposition when heated or exposed to light and air for a few weeks. (9) THF, the solvent used in some of the time-resolved fluorescence measurements, could not be used as the solvent for CD spectra because the UV cutoff of THF is 210 nm.
Electron Transfer Rates in Dichromophoric Peptides
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5279 conditions, and possibly without using high dielectric constant solvents, we prepared oligopeptides 3 and 4, which are
Figure 3. Circular dichroism spectrum of a 1 × 10-4 M solution of 1 in CH3CN/H2O/guanidinium carbonate.
k1/k2, decreased from 27 to 11 (Table 1, entries 1 and 2). Although guanidinium carbonate is not soluble in THF, it is likely that a small amount of salt slowly interacts by intramolecular hydrogen bonding with the helix, probably partially unfolding the peptides. When water was added to THF solutions of 1 and 2, the ratio k1/k2 decreased to 1.8 (Table 1, entry 3). With and without guanidinium, k1/k2 approaches unity for aqueous solutions where helix-coil transition has taken place.10 Finally, the electron transfer rates of 1 and 2 were identical in 2,2,2-trifluoroethanol, the solvent in which the peptides exist as random coils (Table 1, entry 6, and Figure 2). From the data reported in Table 1 it appears that a decrease of k1/k2 is consistently observed when peptides are unfolded (protic solvents) from that observed when the same compounds exist as helices (nonprotic solvents). However, an electric field is smaller in high dielectric constant () solvents, and the decrease of the effective k1/k2 upon addition of water to THF could be the result not only of the helix disruption but also of the increase in dielectric constant of the medium.11 This effect cannot be responsible for the results observed in 2,2,2-trifluoroethanol: k1/k2 in acetonitrile ( ) 37) is 7, and in 2,2,2trifluoroethanol ( ) 27) it is 1.2 (Table 1, entries 7 and 6). Since the peptides exist as R-helices in acetonitrile and as random coils in 2,2,2-trifluoroethanol, Figure 2, the identical electron transfer rates of 1 and 2 in 2,2,2-trifluoroethanol probably result from the cancellation of the helix electric field upon unfolding of the helix. Finally, it is important to emphasize that the observed complex multiexponential decays yield only approximations to microscopic rate constants. Proline-Substituted Peptides. To study electron transfer rates in the absence of helical structure, but otherwise identical (10) The k1/k2 ratio did not change appreciably when the measurements were repeated on the same samples on sequential days (Table 1, entries 4 and 5), implying that the unfolding process had reached equilibrium by the time the measurements were made. (11) Lockhart and Kim found that the effective value of the dielectric constant at the boundary of an oligopeptide was lower than the dielectric constant of the solvent, and the experimentally measured field was 1 order of magnitude stronger than expected on the basis of the of the bulk solvent.3c For calculations of in proteins, see also: (a) King, G; Lee, F. S.; Warshel, A. J. Chem. Phys. 1991, 95, 4366. (b) Wipff, G.; Dearing, A.; Weiner, P. K.; Blaney, J. M.; Kollman, P. A. J. Am. Chem. Soc. 1983, 105, 997. (c) Blaney, J. M.; Weiner, P. K.; Dearing, A.; Kollman, P. A.; Jorgensen, E. C.; Oatley, S. J.; Burridge, J. M.; Blake, C. C. F. J. Am. Chem. Soc. 1982, 104, 6424.
analogous to 1 and 2 except that L-proline has replaced Aib in the backbone.12 In a peptide formed upon coupling proline with another amino acid, there are no amide protons on the proline residue which are available for intramolecular hydrogen bonding. Although oligoproline (L-Pro)n may have a helical structure induced by the five-membered ring constraint,7 the helical structure and the consequent electric field along the helix are perturbed in proline-containing peptides, because the intramolecular hydrogen bonding pattern is disrupted. Lockhart and Kim, for example, employed proline as the helix breaking residue to prepare nonhelical oligopeptides as controls in their studies of the helix electric field.3c Furthermore, the absorption spectra of these proline-containing oligopeptides did not exhibit the characteristic shift caused by the helical peptide electric field. However, the prediction of precise secondary structure of proline-containing peptides is not simple because they can adopt a variety of conformations. The rotational restriction about the C-N-C bonds in the proline five-membered ring limits the rotation about the peptide bond. As a result, proline can adopt cis and trans conformations, which can be interconverted, Figure 4, with an energy barrier of about 16 kcal/mol.13e In (L-Pro)n the cis-trans isomerization equilibrium is affected by the polarity of the solvent, pH, and ionic strength of the medium, the trans conformation being favored in hydrogen-bonding solvents, at low pH, or in the presence of salts such as CaCl2. On the basis of CD7,12c and 1H and 13C NMR13 spectra of the two conformers of oligoproline in water, MeOH, and DMSO, it is thought to be generally true that the all-cis form of (LPro)n is a right-handed helix and the all-trans form of (L-Pro)n is a left-handed, more extended helix.7,12,13 (12) Because of the conformational restraint of the five-membered ring, all-cis- or all-trans-(L-Pro)n peptides are more rigid than an R-helical peptide, and are valuable rigid spacers. Photoinduced electron transfer across oligoproline peptides (0 < n < 4) has been extensively studied by Isied and co-workers. Their models are structurally very different from 3 and 4, as the donor and acceptor (inorganic complexes) are linked at the two ends of the peptide backbones, which contain only proline units (nonalternating), and the experimental conditions (low pH, water) were selected to favor the trans conformation. Therefore, a direct comparison of the conformation with 3 and 4 with elegant earlier work by the Isied group is not possible. (a) Vassilian, A.; Wishart, J. F.; van Hemelryck, B.; Schwarz, H.; Isied, S. S. J. Am. Chem. Soc. 1990, 112, 7278. (b) Isied, S. S.; Vassilian, A.; Magnuson, R. H.; Schwarz, H. J. Am. Chem. Soc. 1985, 107, 7432. (c) Isied, S. S.; Vassilian, A. J. Am. Chem. Soc. 1984, 106, 1732 and references therein for energy-transfer studies across oligoproline. (13) For 1H and 13C NMR studies of the cis-trans equilibrium in oligoproline- and proline-substituted peptides, see: (a) Chao, Y.-Y. H.; Bersohn, R. Biopolymers 1978, 17, 2761. (b) Idem. Biopolymers 1977, 16, 277. (c) Grathwohl, C.; Wu¨thrich, K. Biopolymers 1976, 15, 2025. (d) Ibid. p 2043. For molecular mechanics calculations of preferred conformations of (L-Pro-X)n peptides, see: (a) McDonald, D. Q.; Still, W. C. J. Org. Chem. 1996, 61, 1385. (b) Oka, M.; Nakajima, A. Polym. Bull. 1994, 33, 693.
5280 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Fox and Galoppini
Table 1. Time-Resolved Fluorescence of 1 and 2a,b solvent
1 τ (ns) (A)c
2 τ (ns) (A)c
ket1d (×10-8 s-1)
ket2d (×10-8 s-1)
ket1/ket2
THF THFe THF/H2O THF/H2Oe THF/H2Oe,f CF3CH2OH CH3CN
0.7 (42), 2.5 (58) 1.5 (62), 3.7 (34), 59 (4) 2.5 (53), 8.6 (40), 67 (7) 8 (38), 2.7 (54), 70 (6) 3.7 (61), 14 (31), 97 (8) 16 (83), 46 (9), 162 (8) 0.9 (67), 27 (33)
33 (88), 81 (11) 39 (93), 117 (7) 19 (92), 56 (8) 17 (85), 46 (15) 19 (90), 80 (10) 17 (85), 46 (15) 9 (97), 96 (2)
6.2 2.2 1.0 1.1 0.6 0.3 7.1
0.2 0.2 0.5 0.4 0.3 0.2 0.9
27 11 1.8 2.4 1.8 1.2 8
a The time-resolved fluorescence experiments were performed on a single-photon counting apparatus, with a 20 ps time resolution, λ exc ) 344 nm, λobs ) 400 nm. T ) 20-22 °C. b Concentrations were low (=30 µM, od(355) < 0.1) in order to avoid self-aggregation. Samples were degassed by bubbling Ar for 20 min before and during the measurements. c The numbers in parentheses indicate the relative weighting, A ) amplitude × 100, of the component. d Calculated from ki ) 1/〈τ〉i - 1/τo where 〈τ〉i ) τ′(A′) + τ′′(A′′), where A′ and A′′ are the amplitude of each component and τo ) 300 ns (time decay of a control oligopeptide substituted only with pyrene). e Guanidinium carbonate (about 4 g/mL) added. f Measurement repeated on sample of entry 4 after 1 day.
Figure 4. cis-trans isomerization of proline-substituted peptides.
The conformation of (L-Pro-X)n peptides, where X ) Ala or a non-proline peptide, is more complicated because the equilibrium is dependent also on the nature of X and on the several possible hydrogen bonding patterns between the carbonyl group of Pro and the NH of X.13e Therefore, in addition to cis and trans isomerism, the internal hydrogen bonding pattern in these mixed polypeptides generates β and/or γ turns, and the number of possible conformations is higher than in the parent (L-Pro)n peptides. In most organic solvents, proline-substituted peptides exist as a mixture of conformers and are nonhelical. Oligopeptides 3 and 4 were prepared following the same standard solution-phase coupling procedure employed for the synthesis of 1 and 2.4a,14 Both oligopeptides were much more polar on silica gel and less soluble in organic solvents than 1 and 2, probably because proline-rich peptides are more rigid and extended than alternating Ala-Aib oligomers. The 1H and 13C NMR spectra in CDCl of 3 and 4, as well as the spectra of 3 some of the intermediates, show the presence of more than one conformer.15 The cis-trans distribution in 3 and 4 depends on the pH and solvent polarity, complicating the interpretation of the time-resolved fluorescence measurements. However, from literature data of (L-Pro-L-Ala)n peptides and by analysis of their CD spectra (Vide infra), it appears that the trans conformation of the Pro-Ala bonds is the dominant conformation in the conditions used for the data reported in Table 2.13c,d In the 1H NMR spectrum of 4, one of the aniline doublets is shifted downfield by 0.2 ppm from the corresponding doublet in 3, Figure 5, indicating a different average environment for this group in CDCl3.16 Circular Dichroism Spectra of 3 and 4. An analysis of the CD spectra of 3 and 4 in CH3CN shows that 3 and 4 have a low helical content, as evidenced by the single minimum at approximately 230 nm, Figure 6. The CD spectra of 3 and 4 (14) For general coupling procedures, see: Bodanszky, M. The Practice of Peptide Synthesis; Springer-Verlag: Berlin, 1994. (15) The 1H and 13C NMR spectra in CDCl3 of peptides 3 and 4 and of their precursors are a mixture of at least two isomers, one present in a smaller amount (8-25%). In the Experimental Section, we report the resonances of the major component, but we were not able to unambiguously assign each peak to a cis or trans conformation. NMR spectra of Z-Ala-Pro-AlaOMe in CD3OD, D2O, and DMSO showed that in these solvents the ProAla bonds are mostly trans, with 11-13% of cis.13c (16) No difference in the chemical shift of the chromophore protons was observed between the spectra of 1 and 2.
Figure 5. 1H NMR spectra in CDCl3 of 3 (top) and 4 (bottom) in the aromatic region.
in EtOH and CH3CN, with negative bands at λ ) 210 nm and weak positive bands at 230 nm, are similar in shape to the previously reported CD spectra of all-trans-polyproline and are very different from the CD spectra of all-cis-polyproline, which has a strong negative band at λ ) 190 nm and a strong positive band at λ ) 220 nm.12a Protic solvents also influence the secondary structure of 4, as shown by the CD spectra in CH3CN, EtOH, MeOH, CH3CN/H2O (70/30 (v/v)), and CF3CH2OH, Figure 7.9 As in 1 and 2, peptides 3 and 4 are completely unfolded in the presence of water or in CF3CH2OH. Although 1 and 2 are denatured in protic solvents because of disruption of intramolecular hydrogen bonding, the changes in the secondary structure established by the CD spectra in protic solvents of 3 and 4 may instead reflect changes in the cis-trans equilibrium and/or denaturation caused by interaction of the solvent with the intramolecular hydrogen bonds13e between the proline carbonyl and the alanine NH groups. Steady-State Fluorescence Spectra of 3 and 4. The fluorescence spectra of 3 and 4 in THF, CH2Cl2, MeOH, EtOH, and CH3CN exhibit dramatic solvent dependence, Figure 8. In nonpolar solvents, like THF and CH2Cl2, a broad structureless fluorescence emission is observed at λmax ) 500 nm for 4 and at λmax ) 520 nm for 3. This emission has a lower intensity or is absent in polar or protic solvents such as CH3CN, EtOH, and MeOH, Figure 5. In all solvents studied, the emission intensity of 4 was higher than that of 3. In both 3 and 4, substantial quenching of the pyrene singlet by the N,Ndimethylanilino group can be observed (Φf3 ) 0.13, Φf4 )
Electron Transfer Rates in Dichromophoric Peptides
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5281
Figure 7. Circular dichroism spectra of 1 × 10-4 M solutions of 4 in EtOH (a), CH3CN (b), MeOH (c), CH3CN/H2O (d), and CF3CH2OH (e).
Figure 6. (A) Circular dichroism spectra of 1 × 10-4 M solutions of 1 (a) and 3 (b) in CH3CN. (B) CD spectra of 1 × 10-4 M solutions of 2 (a) and 4 (b) in CH3CN.
0.061, in CH3CN, referenced to pyrene in ethanol, Φfpyr ) 0.7).17 The fluorescence quantum yield of 3 was twice that of 4, indicating more efficient quenching in 4. To establish whether the 500 nm band is an exciplex or excimer emission, and whether it is formed intramolecularly or intermolecularly (peptides 3 and 4 are sparingly soluble in nonpolar solvents and may aggregate), we studied the effect of dilution on the fluorescence spectra of 4 in THF and CH2Cl2, Figure 9. In THF, the emission intensity did not decrease with increasing dilution beyond that expected for Beer’s law absorption changes. In CH2Cl2, the emission of a more dilute sample was stronger than that of a more concentrated sample, and then decreased upon further dilution. These observations suggest that the 500 nm emission is probably an intramolecular exciplex, and that the distance between the chromophores is regulated not only by the number of Ala-Pro residues that separate them (17) The fluorescence quantum yields of 1 and 2 were calculated from the formula reported by Eaton, D. F. Pure Appl. Chem. 1988, 60, 1107. The pyrene fluorescence quantum yield is reported by Birks, J. B. Photophysics of Aromatic Molecules; Wiley & Sons: New York, 1970; p 252.
Figure 8. Steady-state fluorescence spectra of 3 (A) and 4 (B) in CH2Cl2, THF, EtOH, MeOH, and CH3CN (λexc ) 355 nm, 1 × 10-4 M). The λmax of the exciplex emission is 520 nm in 3 and 500 nm in 4.
but also by the backbone conformations. In nonprotic and nonpolar solvents, the Pro-Ala bonds adopt preferentially a cis conformation12,13 and the intramolecular hydrogen bonds are not disrupted; hence, the peptide backbone is less extended and the chromophores are closer in space, permitting the formation of an intramolecular exciplex.18 Furthermore, the consistent observation of a more intense exciplex emission in 4 than in 3, a 10 nm red shift of the (18) From the data available, it is not possible to know whether both compounds have identical conformations in more polar solvents in which the peptide backbone is more extended and the exciplex emission is not observed.
5282 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Fox and Galoppini THF/H2O and CF3CH2OH, where the peptides exist as a random coil, the electron transfer rates of 3 and 4 were identical (Table 2, entries 5 and 6, cf. Figure 7). In both MeOH and EtOH, the rates of electron transfer in 4 are faster than in 3 by about a factor of 2 (Table 2, entries 2-4). However, the CD spectra of 3 and 4 in ethanol show that both peptides have a similar residual secondary structure: in methanol the peptides are completely unfolded, implying that the small difference between k3 and k4 in that solvent probably has a conformational origin. In CH3CN, the k4/k3 ratio is about 3, and the CD spectra show that in this solvent both peptides have a residual secondary structure. The observed rate difference in CH3CN is in the direction opposite to that anticipated from an intramolecular dipole effect. That is, faster rates are observed in 4, in which charge separation takes place parallel to the helix field. However, the fluorescence, CD, and NMR spectra have provided evidence that the backbones in 3 and 4 are conformationally mobile and have a low helical content. Since the peptides populate different backbone conformations, it is likely that any residual difference in electron transfer rates between 3 and 4 can be attributed to a conformational effect rather than an electric field effect. Experimental Section
Figure 9. Steady-state fluorescence spectra of 4 in CH2Cl2 (λexc ) 355 nm, 1 × 10-4 M) (od(355) ) 0.061 (a); 0.014, gain ) 10 (b); 0.099 (c); 0.177 (d)). In THF od(355) ) 0.056 (a); 0.124 (b); 0.012, gain ) 10 (c).
emission in 3 and the chemical shift differences between the 1H NMR spectra of 3 and 4 all suggest that the two peptides populate different average conformations in solutions. Time-Resolved Fluorescence Measurements. The electron transfer rates19 of 3 and 4 were measured only in polar solvents in which the exciplex fluorescence is either not observed or has a low intensity: CH3CN, EtOH, MeOH, THF/H2O (70/30 (v/v)), and CF3CH2OH, Table 2.20 In both 3 and 4, the observed fluorescence decays were biexponential and the electron transfer rates were about 1 order of magnitude slower than in 1 and 2. The slower rates are probably the result of the more extended structure adopted by the proline-rich backbone of 3 and 4 in polar solvents, the distance between chromophores being greater than the distance between chromophores in R-helical 1 and 2 (about 10 Å). In (19) The transient absorption spectra of 3 and 4 show the presence of pyrene radical anion at 500 nm, in direct analogy to that observed in the transient absorption spectra of 1 and 2. (20) A possible way to simplify the operative conformational equilibrium was to use the experimental conditions in which (Pro)n peptides are in an all-trans conformation. However, these conditions (acidic aqueous solution) could not be employed, because at low pH the N,N-dimethylanilino group would be protonated and could no longer act as an electron donor. (21) We attempted the synthesis of oligopeptides analogous to 1 and 2, but substituted with the conformationally restricted amino acids I and II. (We thank Dr. Mark Minton for assistance in the preparation of these compounds.)
Unfortunately, we were unsuccessful in incorporating I or II into an AlaAib backbone, because of solubility problems and/or the difficulty of coupling bulky R,R-disubstituted amino acids. The latter is a problem which is well documented in the literature: Kaminski, Z.; Leplawy, M. T.; Olma, A.; Redlinski, A. In Peptides 1980, Proceedings of the European Peptide Symposium; Brunfeldt, K., Ed.; Scriptor: Copenhagen, 1981; Vol. 16, p 201.
Materials. Anhydrous N,N-dimethylformamide (DMF) and 1,4dioxane were used as received (Aldrich). N-Methylmorpholine (NMM) was dried overnight over 4 Å molecular sieves, fractionally distilled over KOH under Ar, and stored in the dark under Ar. Ethanol was distilled from Mg(OEt)2. Solvents for spectroscopic studies were either distilled or spectral grade and used as received. Commercially available 1-hydroxybenzotriazole trihydrate (HOBT), 1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydrochloride (EDCI), (tert-butoxycarbonyl)-L-alanine (Boc-Ala-OH), L-alanine ethyl ester hydrochloride, and L-proline methyl ester hydrochloride (HCl‚H-Ala-OEt) were used without further purification. (L)-Boc-(p-(N,N-dimethylamino)phenyl)alanine and (L)-pyrenyl-1-alanine methyl ester were prepared according to a previously reported procedure.4a Column chromatography was performed on Merck silica gel 60 (400/600 mesh). Thin-layer chromatography was carried out on 0.25 mm Polygram silica gel plates using 10% ethanolic phosphomolybdic acid, 1% ethanolic ninhydrin, and/or UV light and heat as developing agents. Protected unsubstituted oligopeptides were detected by using a previously described developing reagent.22 Instrumentation and Methods. 1H and 13C NMR spectra were recorded at 500 and 125 MHz, respectively, or at 300 and 75 MHz, respectively, on a General Electric GN-500 or on a Varian Unity Plus spectometer. The solvent was CDCl3. Proton spectra were referenced to TMS, and carbon spectra were referenced to the solvent. Coupling constants are reported in hertz, with a precision of 0.1 Hz. MS and HRMS spectra were obtained on a VG2AB2-E mass spectrometer (VG analytical LDT). Circular dichroism measurements (Figures 2, 3, 6, and 7) were performed on a Jasco J600 spectropolarimeter with a 1 cm quartz cell. The solutions had an optical density
